Esempio n. 1
0
static void destroy_data(args_t *args)
{
    bcf_sr_destroy(args->files);
    int i;
    for (i=0; i<args->vcf_rbuf.m; i++)
        if ( args->vcf_buf[i] ) bcf_destroy1(args->vcf_buf[i]);
    free(args->vcf_buf);
    free(args->fa_buf.s);
    if ( args->mask ) regidx_destroy(args->mask);
    if ( args->chain_fname )
        if ( fclose(args->fp_chain) ) error("Close failed: %s\n", args->chain_fname);
    if ( fclose(args->fp_out) ) error("Close failed: %s\n", args->output_fname);
}
Esempio n. 2
0
int run(int argc, char **argv)
{
    char *trio_samples = NULL, *trio_file = NULL, *rules_fname = NULL, *rules_string = NULL;
    memset(&args,0,sizeof(args_t));
    args.mode = 0;
    args.output_fname = "-";

    static struct option loptions[] =
    {
        {"trio",1,0,'t'},
        {"trio-file",1,0,'T'},
        {"delete",0,0,'d'},
        {"list",1,0,'l'},
        {"count",0,0,'c'},
        {"rules",1,0,'r'},
        {"rules-file",1,0,'R'},
        {"output",required_argument,NULL,'o'},
        {"output-type",required_argument,NULL,'O'},
        {0,0,0,0}
    };
    int c;
    while ((c = getopt_long(argc, argv, "?ht:T:l:cdr:R:o:O:",loptions,NULL)) >= 0)
    {
        switch (c) 
        {
            case 'o': args.output_fname = optarg; break;
            case 'O':
                      switch (optarg[0]) {
                          case 'b': args.output_type = FT_BCF_GZ; break;
                          case 'u': args.output_type = FT_BCF; break;
                          case 'z': args.output_type = FT_VCF_GZ; break;
                          case 'v': args.output_type = FT_VCF; break;
                          default: error("The output type \"%s\" not recognised\n", optarg);
                      };
                      break;
            case 'R': rules_fname = optarg; break;
            case 'r': rules_string = optarg; break;
            case 'd': args.mode |= MODE_DELETE; break;
            case 'c': args.mode |= MODE_COUNT; break;
            case 'l': 
                if ( !strcmp("+",optarg) ) args.mode |= MODE_LIST_GOOD; 
                else if ( !strcmp("x",optarg) ) args.mode |= MODE_LIST_BAD; 
                else error("The argument not recognised: --list %s\n", optarg);
                break;
            case 't': trio_samples = optarg; break;
            case 'T': trio_file = optarg; break;
            case 'h':
            case '?':
            default: error("%s",usage()); break;
        }
    }
    if ( rules_fname )
        args.rules = regidx_init(rules_fname, parse_rules, NULL, sizeof(rule_t), &args);
    else
        args.rules = init_rules(&args, rules_string);
    if ( !args.rules ) return -1;
    args.itr     = regitr_init(args.rules);
    args.itr_ori = regitr_init(args.rules);

    char *fname = NULL;
    if ( optind>=argc || argv[optind][0]=='-' )
    {
        if ( !isatty(fileno((FILE *)stdin)) ) fname = "-";  // reading from stdin
        else error("%s",usage());
    }
    else
        fname = argv[optind];

    if ( !trio_samples && !trio_file ) error("Expected the -t/T option\n");
    if ( !args.mode ) error("Expected one of the -c, -d or -l options\n");
    if ( args.mode&MODE_DELETE && !(args.mode&(MODE_LIST_GOOD|MODE_LIST_BAD)) ) args.mode |= MODE_LIST_GOOD|MODE_LIST_BAD;

    args.sr = bcf_sr_init();
    if ( !bcf_sr_add_reader(args.sr, fname) ) error("Failed to open %s: %s\n", fname,bcf_sr_strerror(args.sr->errnum));
    args.hdr = bcf_sr_get_header(args.sr, 0);
    args.out_fh = hts_open(args.output_fname,hts_bcf_wmode(args.output_type));
    if ( args.out_fh == NULL ) error("Can't write to \"%s\": %s\n", args.output_fname, strerror(errno));
    bcf_hdr_write(args.out_fh, args.hdr);


    int i, n = 0;
    char **list;
    if ( trio_samples )
    {
        args.ntrios = 1;
        args.trios = (trio_t*) calloc(1,sizeof(trio_t));
        list = hts_readlist(trio_samples, 0, &n);
        if ( n!=3 ) error("Expected three sample names with -t\n");
        args.trios[0].imother = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[0]);
        args.trios[0].ifather = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[1]);
        args.trios[0].ichild  = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[2]);
        for (i=0; i<n; i++) free(list[i]);
        free(list);
    }
    if ( trio_file )
    {
        list = hts_readlist(trio_file, 1, &n);
        args.ntrios = n;
        args.trios = (trio_t*) calloc(n,sizeof(trio_t));
        for (i=0; i<n; i++)
        {
            char *ss = list[i], *se;
            se = strchr(ss, ',');
            if ( !se ) error("Could not parse %s: %s\n",trio_file, ss);
            *se = 0;
            args.trios[i].imother = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, ss);
            if ( args.trios[i].imother<0 ) error("No such sample: \"%s\"\n", ss);
            ss = ++se; 
            se = strchr(ss, ',');
            if ( !se ) error("Could not parse %s\n",trio_file);
            *se = 0;
            args.trios[i].ifather = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, ss);
            if ( args.trios[i].ifather<0 ) error("No such sample: \"%s\"\n", ss);
            ss = ++se; 
            if ( *ss=='\0' ) error("Could not parse %s\n",trio_file);
            args.trios[i].ichild = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, ss);
            if ( args.trios[i].ichild<0 ) error("No such sample: \"%s\"\n", ss);
            free(list[i]);
        }
        free(list);
    }

    while ( bcf_sr_next_line(args.sr) )
    {
        bcf1_t *line = bcf_sr_get_line(args.sr,0);
        line = process(line);
        if ( line )
        {
            if ( line->errcode ) error("TODO: Unchecked error (%d), exiting\n",line->errcode);
            bcf_write1(args.out_fh, args.hdr, line);
        }
    }


    fprintf(stderr,"# [1]nOK\t[2]nBad\t[3]nSkipped\t[4]Trio\n");
    for (i=0; i<args.ntrios; i++)
    {
        trio_t *trio = &args.trios[i];
        fprintf(stderr,"%d\t%d\t%d\t%s,%s,%s\n", 
            trio->nok,trio->nbad,args.nrec-(trio->nok+trio->nbad),
            bcf_hdr_int2id(args.hdr, BCF_DT_SAMPLE, trio->imother),
            bcf_hdr_int2id(args.hdr, BCF_DT_SAMPLE, trio->ifather),
            bcf_hdr_int2id(args.hdr, BCF_DT_SAMPLE, trio->ichild)
            );
    }
    free(args.gt_arr);
    free(args.trios);
    regitr_destroy(args.itr);
    regitr_destroy(args.itr_ori);
    regidx_destroy(args.rules);
    bcf_sr_destroy(args.sr);
    if ( hts_close(args.out_fh)!=0 ) error("Error: close failed\n");
    return 0;
}
Esempio n. 3
0
int bam_mpileup(int argc, char *argv[])
{
    int c;
    const char *file_list = NULL;
    char **fn = NULL;
    int nfiles = 0, use_orphan = 0, noref = 0;
    mplp_conf_t mplp;
    memset(&mplp, 0, sizeof(mplp_conf_t));
    mplp.min_baseQ = 13;
    mplp.capQ_thres = 0;
    mplp.max_depth = 250; mplp.max_indel_depth = 250;
    mplp.openQ = 40; mplp.extQ = 20; mplp.tandemQ = 100;
    mplp.min_frac = 0.002; mplp.min_support = 1;
    mplp.flag = MPLP_NO_ORPHAN | MPLP_REALN | MPLP_SMART_OVERLAPS;
    mplp.argc = argc; mplp.argv = argv;
    mplp.rflag_filter = BAM_FUNMAP | BAM_FSECONDARY | BAM_FQCFAIL | BAM_FDUP;
    mplp.output_fname = NULL;
    mplp.output_type = FT_VCF;
    mplp.record_cmd_line = 1;
    mplp.n_threads = 0;
    mplp.bsmpl = bam_smpl_init();

    static const struct option lopts[] =
    {
        {"rf", required_argument, NULL, 1},   // require flag
        {"ff", required_argument, NULL, 2},   // filter flag
        {"incl-flags", required_argument, NULL, 1},
        {"excl-flags", required_argument, NULL, 2},
        {"output", required_argument, NULL, 3},
        {"open-prob", required_argument, NULL, 4},
        {"ignore-RG", no_argument, NULL, 5},
        {"ignore-rg", no_argument, NULL, 5},
        {"gvcf", required_argument, NULL, 'g'},
        {"no-reference", no_argument, NULL, 7},
        {"no-version", no_argument, NULL, 8},
        {"threads",required_argument,NULL,9},
        {"illumina1.3+", no_argument, NULL, '6'},
        {"count-orphans", no_argument, NULL, 'A'},
        {"bam-list", required_argument, NULL, 'b'},
        {"no-BAQ", no_argument, NULL, 'B'},
        {"no-baq", no_argument, NULL, 'B'},
        {"adjust-MQ", required_argument, NULL, 'C'},
        {"adjust-mq", required_argument, NULL, 'C'},
        {"max-depth", required_argument, NULL, 'd'},
        {"redo-BAQ", no_argument, NULL, 'E'},
        {"redo-baq", no_argument, NULL, 'E'},
        {"fasta-ref", required_argument, NULL, 'f'},
        {"read-groups", required_argument, NULL, 'G'},
        {"region", required_argument, NULL, 'r'},
        {"regions", required_argument, NULL, 'r'},
        {"regions-file", required_argument, NULL, 'R'},
        {"targets", required_argument, NULL, 't'},
        {"targets-file", required_argument, NULL, 'T'},
        {"min-MQ", required_argument, NULL, 'q'},
        {"min-mq", required_argument, NULL, 'q'},
        {"min-BQ", required_argument, NULL, 'Q'},
        {"min-bq", required_argument, NULL, 'Q'},
        {"ignore-overlaps", no_argument, NULL, 'x'},
        {"output-type", required_argument, NULL, 'O'},
        {"samples", required_argument, NULL, 's'},
        {"samples-file", required_argument, NULL, 'S'},
        {"annotate", required_argument, NULL, 'a'},
        {"ext-prob", required_argument, NULL, 'e'},
        {"gap-frac", required_argument, NULL, 'F'},
        {"tandem-qual", required_argument, NULL, 'h'},
        {"skip-indels", no_argument, NULL, 'I'},
        {"max-idepth", required_argument, NULL, 'L'},
        {"min-ireads ", required_argument, NULL, 'm'},
        {"per-sample-mF", no_argument, NULL, 'p'},
        {"per-sample-mf", no_argument, NULL, 'p'},
        {"platforms", required_argument, NULL, 'P'},
        {NULL, 0, NULL, 0}
    };
    while ((c = getopt_long(argc, argv, "Ag:f:r:R:q:Q:C:Bd:L:b:P:po:e:h:Im:F:EG:6O:xa:s:S:t:T:",lopts,NULL)) >= 0) {
        switch (c) {
        case 'x': mplp.flag &= ~MPLP_SMART_OVERLAPS; break;
        case  1 :
            mplp.rflag_require = bam_str2flag(optarg);
            if ( mplp.rflag_require<0 ) { fprintf(stderr,"Could not parse --rf %s\n", optarg); return 1; }
            break;
        case  2 :
            mplp.rflag_filter = bam_str2flag(optarg);
            if ( mplp.rflag_filter<0 ) { fprintf(stderr,"Could not parse --ff %s\n", optarg); return 1; }
            break;
        case  3 : mplp.output_fname = optarg; break;
        case  4 : mplp.openQ = atoi(optarg); break;
        case  5 : bam_smpl_ignore_readgroups(mplp.bsmpl); break;
        case 'g':
            mplp.gvcf = gvcf_init(optarg);
            if ( !mplp.gvcf ) error("Could not parse: --gvcf %s\n", optarg);
            break;
        case 'f':
            mplp.fai = fai_load(optarg);
            if (mplp.fai == NULL) return 1;
            mplp.fai_fname = optarg;
            break;
        case  7 : noref = 1; break;
        case  8 : mplp.record_cmd_line = 0; break;
        case  9 : mplp.n_threads = strtol(optarg, 0, 0); break;
        case 'd': mplp.max_depth = atoi(optarg); break;
        case 'r': mplp.reg_fname = strdup(optarg); break;
        case 'R': mplp.reg_fname = strdup(optarg); mplp.reg_is_file = 1; break;
        case 't':
                  // In the original version the whole BAM was streamed which is inefficient
                  //  with few BED intervals and big BAMs. Todo: devise a heuristic to determine
                  //  best strategy, that is streaming or jumping.
                  if ( optarg[0]=='^' ) optarg++;
                  else mplp.bed_logic = 1;
                  mplp.bed = regidx_init(NULL,regidx_parse_reg,NULL,0,NULL);
                  mplp.bed_itr = regitr_init(mplp.bed);
                  if ( regidx_insert_list(mplp.bed,optarg,',') !=0 )
                  {
                      fprintf(stderr,"Could not parse the targets: %s\n", optarg);
                      exit(EXIT_FAILURE);
                  }
                  break;
        case 'T':
                  if ( optarg[0]=='^' ) optarg++;
                  else mplp.bed_logic = 1;
                  mplp.bed = regidx_init(optarg,NULL,NULL,0,NULL);
                  if (!mplp.bed) { fprintf(stderr, "bcftools mpileup: Could not read file \"%s\"", optarg); return 1; }
                  break;
        case 'P': mplp.pl_list = strdup(optarg); break;
        case 'p': mplp.flag |= MPLP_PER_SAMPLE; break;
        case 'B': mplp.flag &= ~MPLP_REALN; break;
        case 'I': mplp.flag |= MPLP_NO_INDEL; break;
        case 'E': mplp.flag |= MPLP_REDO_BAQ; break;
        case '6': mplp.flag |= MPLP_ILLUMINA13; break;
        case 's': if ( bam_smpl_add_samples(mplp.bsmpl,optarg,0)<0 ) error("Could not read samples: %s\n",optarg); break;
        case 'S': if ( bam_smpl_add_samples(mplp.bsmpl,optarg,1)<0 ) error("Could not read samples: %s\n",optarg); break;
        case 'O': 
            switch (optarg[0]) {
                case 'b': mplp.output_type = FT_BCF_GZ; break;
                case 'u': mplp.output_type = FT_BCF; break;
                case 'z': mplp.output_type = FT_VCF_GZ; break;
                case 'v': mplp.output_type = FT_VCF; break;
                default: error("[error] The option \"-O\" changed meaning when mpileup moved to bcftools. Did you mean: \"bcftools mpileup --output-type\" or \"samtools mpileup --output-BP\"?\n"); 
            }
            break;
        case 'C': mplp.capQ_thres = atoi(optarg); break;
        case 'q': mplp.min_mq = atoi(optarg); break;
        case 'Q': mplp.min_baseQ = atoi(optarg); break;
        case 'b': file_list = optarg; break;
        case 'o': {
                char *end;
                long value = strtol(optarg, &end, 10);
                // Distinguish between -o INT and -o FILE (a bit of a hack!)
                if (*end == '\0') mplp.openQ = value;
                else mplp.output_fname = optarg;
            }
            break;
        case 'e': mplp.extQ = atoi(optarg); break;
        case 'h': mplp.tandemQ = atoi(optarg); break;
        case 'A': use_orphan = 1; break;
        case 'F': mplp.min_frac = atof(optarg); break;
        case 'm': mplp.min_support = atoi(optarg); break;
        case 'L': mplp.max_indel_depth = atoi(optarg); break;
        case 'G': bam_smpl_add_readgroups(mplp.bsmpl, optarg, 1); break;
        case 'a':
            if (optarg[0]=='?') {
                list_annotations(stderr);
                return 1;
            }
            mplp.fmt_flag |= parse_format_flag(optarg);
        break;
        default:
            fprintf(stderr,"Invalid option: '%c'\n", c);
            return 1;
        }
    }

    if ( mplp.gvcf && !(mplp.fmt_flag&B2B_FMT_DP) )
    {
        fprintf(stderr,"[warning] The -t DP option is required with --gvcf, switching on.\n");
        mplp.fmt_flag |= B2B_FMT_DP;
    }
    if ( mplp.flag&(MPLP_BCF|MPLP_VCF|MPLP_NO_COMP) )
    {
        if ( mplp.flag&MPLP_VCF )
        {
            if ( mplp.flag&MPLP_NO_COMP ) mplp.output_type = FT_VCF;
            else mplp.output_type = FT_VCF_GZ;
        }
        else if ( mplp.flag&MPLP_BCF )
        {
            if ( mplp.flag&MPLP_NO_COMP ) mplp.output_type = FT_BCF;
            else mplp.output_type = FT_BCF_GZ;
        }
    }
    if ( !(mplp.flag&MPLP_REALN) && mplp.flag&MPLP_REDO_BAQ )
    {
        fprintf(stderr,"Error: The -B option cannot be combined with -E\n");
        return 1;
    }
    if (use_orphan) mplp.flag &= ~MPLP_NO_ORPHAN;
    if (argc == 1)
    {
        print_usage(stderr, &mplp);
        return 1;
    }
    if (!mplp.fai && !noref) {
        fprintf(stderr,"Error: mpileup requires the --fasta-ref option by default; use --no-reference to run without a fasta reference\n");
        return 1;
    }
    int ret,i;
    if (file_list) 
    {
        if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
        mplp.files  = fn;
        mplp.nfiles = nfiles;
    }
    else
    {
        mplp.nfiles = argc - optind;
        mplp.files  = (char**) malloc(mplp.nfiles*sizeof(char*));
        for (i=0; i<mplp.nfiles; i++) mplp.files[i] = strdup(argv[optind+i]);
    }
    ret = mpileup(&mplp);

    for (i=0; i<mplp.nfiles; i++) free(mplp.files[i]);
    free(mplp.files);
    free(mplp.reg_fname); free(mplp.pl_list);
    if (mplp.fai) fai_destroy(mplp.fai);
    if (mplp.bed) regidx_destroy(mplp.bed);
    if (mplp.bed_itr) regitr_destroy(mplp.bed_itr);
    if (mplp.reg) regidx_destroy(mplp.reg);
    bam_smpl_destroy(mplp.bsmpl);
    return ret;
}
Esempio n. 4
0
File: tabix.c Progetto: Illumina/akt
static int query_regions(args_t *args, char *fname, char **regs, int nregs)
{
    int i;
    htsFile *fp = hts_open(fname,"r");
    if ( !fp ) error("Could not read %s\n", fname);
    enum htsExactFormat format = hts_get_format(fp)->format;

    regidx_t *reg_idx = NULL;
    if ( args->targets_fname )
    {
        reg_idx = regidx_init(args->targets_fname, NULL, NULL, 0, NULL);
        if ( !reg_idx ) error("Could not read %s\n", args->targets_fname);
    }

    if ( format == bcf )
    {
        htsFile *out = hts_open("-","w");
        if ( !out ) error("Could not open stdout\n", fname);
        hts_idx_t *idx = bcf_index_load(fname);
        if ( !idx ) error("Could not load .csi index of %s\n", fname);
        bcf_hdr_t *hdr = bcf_hdr_read(fp);
        if ( !hdr ) error("Could not read the header: %s\n", fname);
        if ( args->print_header )
            bcf_hdr_write(out,hdr);
        if ( !args->header_only )
        {
            bcf1_t *rec = bcf_init();
            for (i=0; i<nregs; i++)
            {
                hts_itr_t *itr = bcf_itr_querys(idx,hdr,regs[i]);
                while ( bcf_itr_next(fp, itr, rec) >=0 )
                {
                    if ( reg_idx && !regidx_overlap(reg_idx, bcf_seqname(hdr,rec),rec->pos,rec->pos+rec->rlen-1, NULL) ) continue;
                    bcf_write(out,hdr,rec);
                }
                tbx_itr_destroy(itr);
            }
            bcf_destroy(rec);
        }
        if ( hts_close(out) ) error("hts_close returned non-zero status for stdout\n");
        bcf_hdr_destroy(hdr);
        hts_idx_destroy(idx);
    }
    else if ( format==vcf || format==sam || format==unknown_format )
    {
        tbx_t *tbx = tbx_index_load(fname);
        if ( !tbx ) error("Could not load .tbi/.csi index of %s\n", fname);
        kstring_t str = {0,0,0};
        if ( args->print_header )
        {
            while ( hts_getline(fp, KS_SEP_LINE, &str) >= 0 )
            {
                if ( !str.l || str.s[0]!=tbx->conf.meta_char ) break;
                puts(str.s);
            }
        }
        if ( !args->header_only )
        {
            int nseq;
            const char **seq = NULL;
            if ( reg_idx ) seq = tbx_seqnames(tbx, &nseq);
            for (i=0; i<nregs; i++)
            {
                hts_itr_t *itr = tbx_itr_querys(tbx, regs[i]);
                if ( !itr ) continue;
                while (tbx_itr_next(fp, tbx, itr, &str) >= 0)
                {
                    if ( reg_idx && !regidx_overlap(reg_idx,seq[itr->curr_tid],itr->curr_beg,itr->curr_end, NULL) ) continue;
                    puts(str.s);
                }
                tbx_itr_destroy(itr);
            }
            free(seq);
        }
        free(str.s);
        tbx_destroy(tbx);
    }
    else if ( format==bam )
        error("Please use \"samtools view\" for querying BAM files.\n");

    if ( reg_idx ) regidx_destroy(reg_idx);
    if ( hts_close(fp) ) error("hts_close returned non-zero status: %s\n", fname);

    for (i=0; i<nregs; i++) free(regs[i]);
    free(regs);
    return 0;
}